Evaluating and implementing block jackknife resampling Mendelian randomization to mitigate bias induced by overlapping samples

Participant overlap can induce overfitting bias into Mendelian randomization (MR) and polygenic risk score (PRS) studies. Here, we evaluated a block jackknife resampling framework for genome-wide association studies (GWAS) and PRS construction to mitigate overfitting bias in MR analyses and implemented this study design in a causal inference setting using data from the UK Biobank. We simulated PRS and MR under three scenarios: (1) using weighted SNP estimates from an external GWAS, (2) using weighted SNP estimates from an overlapping GWAS sample and (3) using a block jackknife resampling framework. Based on a P-value threshold to derive genetic instruments for MR studies (P < 5 × 10(−8)) and a 10% variance in the exposure explained by all SNPs, block-jackknifing PRS did not suffer from overfitting bias (mean R(2) = 0.034) compared with the externally weighted PRS (mean R(2) = 0.040). In contrast, genetic instruments derived from overlapping samples explained a higher variance (mean R(2) = 0.048) compared with the externally derived score. Overfitting became considerably more severe when using a more liberal P-value threshold to construct PRS (e.g. P < 0.05, overlapping sample PRS mean R(2) = 0.103, externally weighted PRS mean R(2) = 0.086), whereas estimates using jackknife score remained robust to overfitting (mean R(2) = 0.084). Using block jackknife resampling MR in an applied analysis, we examined the effects of body mass index on circulating biomarkers which provided comparable estimates to an externally weighted instrument, whereas the overfitted scores typically provided narrower confidence intervals. Furthermore, we extended this framework into sex-stratified, multivariate and bidirectional settings to investigate the effect of childhood body size on adult testosterone levels.